在携号转网背景下基于Xgboost算法的高价值客户流失预警建模

Under the background of carrying number to transfer network, operators will face the risk of customer loss. It is of great significance to select the high-risk groups with loss tendency from the total number of customers and care and maintain these customers in advance. Firstly, this paper constructs the value model of customer churn early warning and accurate retention; secondly, based on the historical behavior data of customers, it constructs the high-value customer churn early warning model by using Xgboost algorithm; finally, combining the value model and churn early warning model, it gives the best retention strategy to deal with the high value customer churn. The model is applied to the production test environment, showing a good prediction accuracy, and can significantly improve the economic benefits of enterprises.